Abstract
Background:
Mycophenolate mofetil (MMF) is an immunosuppressant commonly used to treat systemic lupus erythematosus (SLE) and lupus nephritis. MMF is a known teratogen associated with significant toxicities, including an increased risk of infections and malignancies. MMF withdrawal is desirable once disease quiescence is achieved, but if and when to do so has not been well-studied.
Methods:
This multi-center, open-label, randomized trial investigates the risks associated with MMF withdrawal compared to MMF maintenance in participants with quiescent (clinical SLEDAI <4) SLE on long-term MMF (mean 6.6 years). The withdrawal group tapered off MMF over 12 weeks, while the maintenance group maintained a baseline dose for 60 weeks. Clinically significant disease reactivation (CSDR) requiring increased doses or new immunosuppressive therapy was the primary endpoint.
Findings:
Of the 100 participants included in the intention-to-treat analysis (49 maintenance, 51 withdrawal), 84% (n=84) were female, 40% (n=40) were White, 41% (n=41) were Black, and 76% (n=76) had a history of lupus nephritis. The average age was 42 (SD, 12.7). By week 60, 17·6% (n=9) of patients in the withdrawal group had CSDR, compared to 10·2% (n=5) in the maintenance group. The risk of CSDR was 11% (95% confidence interval [CI], 5-24%) in the maintenance group and 18% (95% CI, 10-32%) in the withdrawal group. The estimated increase in the risk of CSDR with MMF withdrawal was 7% (one-sided 85% confidence limit [CL], 15%). Similar adverse event rates were observed in the maintenance (45 [90%]) and withdrawal (46 [88.5%]) groups. Infections were more frequent with MMF maintenance (32 [64%]) compared to withdrawal (24 [46%].
Interpretations:
MMF withdrawal is not significantly inferior to MMF maintenance. Estimates for the rates of disease reactivation and increases in risk with withdrawal can assist clinicians in making informed decisions on withdrawing MMF in stable patients. (ClinicalTrials.gov, NCT01946880).
Funding:
The National Institute of Allergy and Infectious Diseases and the National Institute of Arthritis and Musculoskeletal and Skin Diseases.
Keywords: Lupus, lupus nephritis, mycophenolate mofetil, flare, withdrawal
Introduction:
Systemic lupus erythematosus (SLE) is an incurable, chronic autoimmune disease and a leading cause of death for young women in the United States, especially among Black and Hispanic women1. Prolonged disease quiescence is achievable for many with immunosuppressive medications, but associated toxicities contribute to increased morbidity and early mortality, and rates of breakthrough flares are unquantified. The appropriate duration of extended immunosuppressive medication use after disease stabilization remains unknown.
Mycophenolate mofetil (MMF) is an effective treatment for many aspects of moderate/severe or organ-threatening disease activity in SLE, including lupus nephritis.2 MMF treatment, however, requires frequent monitoring for bone marrow suppression and transaminitis, is associated with malaise and gastrointestinal discomfort, and may impair responses to SARS-CoV-2 vaccination.3 The medication label also carries a boxed warning for increased risks of infections and malignancies.4 Furthermore, MMF use during pregnancy is associated with severe embryopathy, resulting in the establishment of a Risk Evaluation and Mitigation Strategy (REMS) program.5 Because of the significant toxicities and morbidities associated with long-term use, MMF withdrawal is desirable once disease quiescence has been achieved; however, if and when to do so has not been well-studied.6,7 This study aimed to determine the effects of MMF withdrawal on the risk of clinically significant disease reactivation (CSDR) in patients with quiescent (clinical SLEDAI <4) SLE on long-term MMF therapy. Here we report the estimated risks of disease reactivation and flare through 60 weeks and the increase in risk with withdrawal. These data may help inform considerations of the risks and benefits of MMF withdrawal and guide decisions regarding MMF discontinuation in quiescent SLE patients.
Methods:
Trial design and oversight:
This open-label, randomized trial of MMF withdrawal in clinically quiescent SLE (NCT01946880) was funded by the Autoimmunity Centers of Excellence Network and sponsored by the National Institute of Allergy and Infectious Diseases at the National Institutes of Health (NIAID/NIH).
Participants were enrolled in 19 US centers (appendix p3). The trial protocol was approved by institutional review boards at each site. An IND exemption was obtained from the US Food and Drug Administration. All participants provided prior written informed consent. The study was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice guidelines. An independent Data and Safety Monitoring Committee regularly reviewed study progress and safety data. Rho Federal Systems Division, Inc. (Durham, NC) held and analyzed the data; data are accessible from ImmPort (www.immport.org).
Eligibility criteria:
Eligible participants were 18-70 years old, met American College of Rheumatology 1997 SLE criteria,8 and had a clinical SLEDAI (modified-SLE Disease Activity Index excluding complement and anti-double-stranded DNA [dsDNA])9 score <4 at screening. MMF therapy was stable or decreasing for ≥2 years if initiated for renal indications or ≥1 year for non-renal indications. Doses of MMF (1-3g/day) and hydroxychloroquine (HCQ) were stable for ≥12 weeks prior to randomization. Prednisone (≤10mg/day) was permitted for the 12 weeks before randomization (with up to four days of temporary use of prednisone [up to 20mg/day] in weeks 12 to four before randomization). MMF dose was stable and disease met the eligibility criteria at screening. Exclusion criteria included other immunosuppressant (including but not limited to azathioprine, methotrexate, 6-mercaptopurine, leflunomide, calcineurin inhibitors, anti-tumor necrosis factor agents), B cell-depleting therapies, belimumab, or cyclophosphamide use, prednisone >25mg/day within 24 weeks, spot urine protein creatinine ratio (UPCR) >1·0mg/mg, serum creatinine >2·0mg/dL, leukocytes <2000/mm3, or platelets <75,000/mm3. Contraception, following REMs criteria, was required.5
Trial Procedures:
Adaptive randomization ensured groups were balanced on key characteristics: study site, renal vs. non-renal disease, and baseline MMF dose (≥2g/day vs. <2g/day).10 The maintenance group continued their baseline MMF dose (1-3g/day). The withdrawal group tapered off MMF over 12 weeks following a structured taper based on baseline dose.
Visits at baseline, monthly through week 24, and at weeks 32, 40, 48, and 60 included standardized SLE disease assessments (Safety of Estrogens in Lupus Erythematosus National Assessment-SLEDAI (SELENA-SLEDAI),9 BILAG 2004 Index (BILAG),11 and the SELENA-SLEDAI flare index12), physical examination findings, medications, and grade ≥2 adverse events (AEs), based on the National Cancer Institute—Common Terminology Criteria for Adverse Events (NCI-CTCAE v4·0).13 Standard laboratory studies included complete blood count, serum creatinine, liver enzymes, dsDNA autoantibody levels, complement C3 and C4, urinalysis, and UPCR.
Study endpoints:
The primary endpoint, probability of CSDR by week 60, was defined as any SELENA-SLEDAI flare with a sustained increase in immunosuppressive therapy as follows: (1) prednisone increased to ≥15mg/day for ≥4 weeks, (2) ≥2 short steroid bursts, or (3) resumption/increased dose of MMF or other immunosuppressants.
Secondary disease activity endpoints were probability-of- and time-to-SELENA-SLEDAI severe or mild/moderate flares and BILAG A (severe) or B (moderate) flares. Changes in the Systemic Lupus International Collaborating Clinics/American College of Rheumatology Disease Damage Index (SLICC/ACR DI) score,14 quality of life measures (Functional Assessment of Chronic Illness fatigue score,15 SF-36,16 and Lupus Quality of Life scores17), AEs, serious AEs (SAEs), and infections were recorded. Mechanistic and time to recovery outcomes have not been analyzed and will be included in a future manuscript.
Statistical analyses and power calculations:
This study evaluated whether the increased risk of disease reactivation associated with MMF withdrawal compared to MMF maintenance was within an acceptable limit under a non-inferiority design using an estimation-based approach. Initially, a panel of 11 rheumatologists and clinical trialists from medical centers in the US, the pharmaceutical industry, and the NIH concluded that the definition of an acceptable limit depends on individual circumstances and physician judgment. In some cases, an increase in CSDR risk of 15% or higher could be acceptable, but wide applicability would require a smaller increase in risk and necessitate a large sample size. For example, at least 796 participants would be needed to demonstrate that risk increases ≤ 10% (85% power, one-sided α=0·05 for non-inferiority, see appendix pp 1-2). Given the very large sample size requirements to test the hypothesis of non-inferiority, we opted for a more feasible estimation-based design to evaluate the absolute increases in risk associated with MMF withdrawal.
A sample size of 120 participants randomized 1:1 was selected to provide reasonable precision estimating the absolute increase in the risk of CSDR by week 60 with MMF withdrawal, under a two-sample comparison of a binary variable. The upper bound to a one-sided 85% confidence limit (CL) was chosen as a reasonable level of precision for determining sample size. Assuming that the real CSDR risk by 60 weeks is 5% with MMF maintenance and 10% with MMF withdrawal, a sample of 120 participants provides an 85% chance that the upper CL for the observed absolute increase in risk will be ≤15%.
The event-free probability of CSDR, and other flare endpoints, over 60 weeks was calculated using the Kaplan-Meier product limit estimator. Confidence intervals were computed using Greenwood’s formula for standard error. All within-group point estimates are presented with two-sided 95% confidence intervals. Estimated increases in risk with withdrawal are presented with one-sided upper 75%, 85%, and 95% CLs. Analyses were repeated using the subset of participants who initiated MMF for lupus nephritis and using female participants given the observed imbalance. To evaluate cumulative disease burden, we performed a post hoc analysis of the mean SLEDAI area under the curve between groups.18 Risk factors for CSDR were explored post-hoc using a Wilcoxon rank sum test or Fisher’s Exact test without correcting for multiple comparisons. Descriptive statistics summarize demographic and baseline characteristics. Analyses of disease activity were conducted in the modified intention-to-treat (mITT) population, defined as all randomized participants who met study entry criteria and began study-provided MMF. Analyses of AEs Kaplan-Meier curves of time to first infection were conducted in the safety population, defined as all participants who initiated study-provided MMF. Analyses were performed using SAS 9.4 (SAS Institute; Cary, NC).
Role of the funding source:
The sponsor, DAIT/NIAID/NIH, provided oversight, coordination and financial support for the trial. The investigators designed the trial and collected and analyzed the data in consultation with the sponsor and the statistical and data management center (Rho Federal Systems). DAIT/NIAID/NIH was involved in trial data interpretation, writing of the manuscript and, along with the academic investigators, made the decision to submit the manuscript for publication
Results:
Between November 2013 and April 2018, 123 participants were screened; 21 failed, and 102 were randomized and treated (safety population: 50 maintenance, 52 withdrawal). Two individuals were later found ineligible (mITT population: 49 maintenance, 51 withdrawal; Figure 1). In the mITT population, 88% (43/49) of maintenance participants completed the 60-week study compared with 94% (48/51) of withdrawal participants.
Figure 1:
Trial Profile. One hundred twenty-three individuals initially consented to participate in the study, and 100 individuals were analyzed as part of the mITT population.
Baseline characteristics were similar between groups a (Table 1). Most participants were women (84%; n=84), with good representation across races and ethnicities. The mean age was 42 years, with an average disease duration of 13 years and an MMF treatment duration of 6·6 years. Most participants had a history of lupus nephritis (76%; n=76), and (60%; n=60) were positive for anti-dsDNA autoantibodies at enrollment, demonstrating serologic activity; however, SELENA-SLEDAI scores were low (mean 2·2; SD, 1·8). Prednisone was used at baseline in 37% (n=37) of participants with an average 4 mg/day dose.
Table 1:
Demographic and baseline characteristics of the mITT population.
| Maintenance (n=49) | Withdrawal (n=51) | Total (n=100) | |
|---|---|---|---|
| Sex, n (%) | 38 (78) | 46 (90) | 84 (84) |
| Female | 38 (78) | 46 (90) | 84 (84) |
| Male | 11 (22) | 5 (10) | 16 (16) |
| Race, n (%) | |||
| White | 24 (49·0) | 16 (31·4) | 40 (40·0) |
| Black | 19 (38·8) | 22 (43·1) | 41 (41·0) |
| Asian | 2 (4·1) | 8 (15·7) | 10 (10·0) |
| American Indian/Alaska Native | 2 (4·1) | 0 | 2 (2·0) |
| Other | 1 (2·0) | 3 (5·9) | 4 (4·0) |
| Unknown | 1 (2·0) | 2 (3·9) | 3 (3·0) |
| Not Hispanic or Latino, n (%) | 39 (79·6) | 40 (78·4) | 79 (79·0) |
| Hispanic/Latino, n (%) | 10 (20) | 11 (22) | 21 (21) |
| White | 8 (16) | 8 (16) | 16 (16) |
| Black | 0 | 1 (2) | 1 (1) |
| Age, years, mean (SD) | 42·2 (12·96) | 41·8 (12·48) | 42·0 (12·66) |
| Weight, kg, mean (SD) | 81·6 (24·60) | 80·3 (18·54) | 81·0 (21·65) |
| Disease duration, years, mean (SD) | 13·7 (8·2) | 12·0 (7·9) | 12·9 (8·0) |
| History of Lupus nephritis, n (%) | 39 (80) | 37 (73) | 76 (76) |
| Prednisone at baseline, n (%) | 19 (38·8) | 18 (35·3) | 37 (37) |
| Dose, mg, mean (SD)* | 4·8 (2·7) | 3·4 (1·8) | 4·1 (2·3) |
| ≤ 5 mg* | 16 (32·3) | 18 (35·3) | 34 (34) |
| >5 -10 mg* | 3 (6·1) | 0 | 3 (3) |
| MMF duration, years, mean (SD) | 6·9 (4·3) | 6·3 (4·2) | 6·6 (4·2) |
| Median (min, max) | 5·4 (1·1, 17·1) | 4·6 (1·7, 16·6) | 5·3 (1·1, 17·1) |
| (Q1, Q3) | 3·4, 10·1 | 2·7, 8·5 | 3·3, 8·8 |
| Baseline MMF Dose, g, mean | 1·604 | 1·642 | 1·623 |
| History of lupus nephritis | 1·590 | 1·723 | 1·655 |
| No history of lupus nephritis | 1·658 | 1·429 | 1·524 |
| Baseline MMF Dose, n (%) | |||
| 1 g | 20 (40·8) | 18 (35·3) | 38 (38) |
| 1 – 2 g | 6 (12·2) | 12 (23·5) | 18 (18) |
| >2 g | 23 (46·9) | 21 (41·2) | 44 (44) |
| SELENA-SLEDAI, mean (SD)** | 2·4 (1·76) | 1·9 (1·75) | 2·2 (1·76) |
| Median (min, max) | 2·0 (0·0, 8·0) | 2·0 (0·0, 6·0) | 2·0 (0·0, 8·0) |
| Any BILAG B, n (%) | 1 (2) | 2 (4) | 3 (3) |
| Met DORIS Remission at screening, n (%) | 40 (82) | 40 (78) | 80 (80) |
| Positive anti-dsDNA, n (%) | 34 (69) | 26 (51) | 60 (60) |
| Low C3 complement, n (%) | 13 (27) | 9 (18) | 22 (22) |
| Low C4 complement, n (%) | 6 (12) | 5 (10) | 11 (11) |
| Serum creatinine, umol/L, mean (SD) | 79·2 (25·22) | 80·7 (29·27) | 80·0 (27·25) |
| Spot urine protein creatinine ratio, mg:mg, mean (SD) | 0·2 (0·26) | 0·2 (0·16) | 0·2 (0·22) |
| Chronic Kidney Disease Stage, n (%) | |||
| Stage 3a | 2 (4.0) | 6 (12) | 8 (8) |
| Stage 3b | 2 (4.0) | 5 (10) | 7 (7) |
| Stage 4 | 0 | 0 | 0 |
Among those taking prednisone at baseline;
Complete SELENA-SLEDAI at baseline, including serologies
SELENA-SLEDAI: Safety of Estrogen in Lupus Erythematosus National Assessment Systemic Lupus Erythematosus Disease Activity Index
By week 60, 14 participants had CSDR, 9 in the withdrawal group (17·6%) and 5 in the maintenance group (10·2%) (Figure 2A and appendix p 4). Twenty-five patients in the withdrawal group (49.0%) and 20 in the maintenance group (40.8%) experienced any SELENA-SLEDAI flare by week 60, while 8 in the withdrawal (15.7%) and 4 in the maintenance (8.2%) experienced a severe SELENA-SLEDAI flare (Figure 1B and Table S2). Similarly, 4 participants in the withdrawal group (7.8%) and 1 in the maintenance group (2.0%) had a BILAG A flare and 25 in the withdrawal group (49.0%) and 21 in the maintenance group (42.9) had either a BILAG A or B flare by week 60 (Figure 2C–D and appendix p 4).
Figure 2:
Kaplan-Meier curves for disease flares in the mITT population (n=100 participants). The y-axis gives the probability of survival without reaching the endpoint. A: Kaplan-Meier curve of time to clinically significant disease reactivation (CSDR) by treatment allocation. B: Curve of time to any SELENA-SLEDAI flare by treatment allocation as measured by the SELENA-SLEDAI flare index. C: Curve of time to BILAG A flares by treatment allocation. D: Time to achieve a BILAG A or B flare by treatment allocation. Numbers indicate the number of patients who did not meet the endpoint while numbers in parenthesis indicate the number of censored patients.
Furthermore, the estimated risk of CSDR by week 60 was 11% (95% CI, 5-24%) in the maintenance group and 18% (95% CI, 10-32%) in the withdrawal group (Figure 3A and appendix p 4). Mean time to CSDR was 38·0 weeks and 38·5 weeks, respectively (appendix p 4). Risk estimates for all measures of disease activity overlapped between the maintenance and withdrawal arms (Figure 3A and appendix p 4) and were consistent in female participants (appendix pp 5 and 13). Three individuals with CSDR (two maintenance; one withdrawal) resumed/increased MMF or received two bursts of steroids but did not meet SELENA-SLEDAI flare criteria based on disease activity alone (appendix pp 6-8). Although not statistically significant, the increase in risk with MMF withdrawal was 6-8% for the CSDR and flare endpoints, with one-sided 85% CLs ranging from 11-19% (Figure 3B). The cumulative glucocorticoid dose over time was not different between groups (appendix p 9). SLICC/ACRDI and quality of life measures in both groups were steady over time (appendix pp 14-18). SELENA-SLEDAI at baseline was low, and the area under the curve remained low over time in both groups (appendix p 10). Post-hoc analyses revealed that low C3 and C4, DORIS remission criteria, and low lymphocyte counts at baseline were possible risk factors for CSDR (appendix p 11).
Figure 3:
Risks of clinically significant disease reactivation (CSDR) and flare endpoints at week 60 with MMF withdrawal. A. Kaplan-Meier estimated risks at week 60 with 95% CI in the mITT population (n=100). B. Observed increase in risk with MMF withdrawal (black star) and one-sided, upper 75, 85, and 95% confidence limits (flat end of the arrow) in the mITT population. C. Kaplan-Meier estimated risks at week 60 with 95% CI in the subset with a history of renal disease (n=76). D. Observed increase in risk with MMF withdrawal (black star) and one-sided, upper 75, 85, and 95% confidence limits (flat end of the arrow) in the subset with a history of renal disease.
Eleven of 76 participants (3 of 39 maintenance, 8 of 37 withdrawal) with a history of renal disease experienced CSDR by 60 weeks, with similar prevalence compared to the mITT population across all endpoints (appendix p 12 and 19). Observed risks were 8% (95% CI, 3-24%) and 23% (95% CI, 12-40%), respectively (Figure 3C and appendix p 12). For other flare endpoints, point estimates and one-sided CLs for increases in risk with withdrawal are higher in the subset with renal disease compared to the mITT analysis set (Figure 3C and appendix p 12). The observed increase in CSDR risk with MMF withdrawal was 14%, with 85% confidence that the actual increase is 23% or less (Figure 3D).
Ninety-one participants experienced AEs (45 [90%] maintenance; 46 [88·5%] withdrawal; Table 2). In both groups, approximately 90% of AEs were grade 2, with similar numbers of grade 3 or 4 AEs (20 in 12 maintenance participants; 15 in 11 withdrawal participants). Twelve participants experienced SAEs (twelve in seven maintenance participants; six in five withdrawal participants). One pregnancy occurred nine months after complete MMF withdrawal, resulting in a full-term, healthy infant. One in each group suffered breast cancer, and one in the maintenance group had a colon adenoma. No deaths occurred.
Table 2:
Adverse events in the safety population.
| Maintenance (n=50)1, 2 | Withdrawal (n=52)1, 2 | |||
|---|---|---|---|---|
|
|
||||
| n (%) | Events, n | n (%) | Events, n | |
| Serious Adverse Events | 7 (14·0) | 12 | 5 (9·6) | 6 |
| Total Adverse Events | 45 (90·0) | 189 | 46 (88·5) | 205 |
| Related to SLE* | 25 (50·0) | 63 | 32 (61·5) | 79 |
| Adverse Events by Severity | ||||
| Grade 1 | 0 | 0 | 1 (1·9) | 1 |
| Grade 2 | 45 (90·0) | 169 | 45 (86·5) | 189 |
| Grade 3 | 10 (20·0) | 18 | 11 (21·2) | 15 |
| Grade 4 | 2 (4·0) | 2 | 0 | 0 |
| Total Infections | 32 (64·0) | 63 | 24 (46·2) | 49 |
| Infections by Severity | ||||
| Grade 2 | 30 (60·0) | 57 | 24 (46·2) | 48 |
| Grade 3 | 4 (8·0) | 6 | 1 (1·9) | 1 |
| Grade 4 | 0 | 0 | 0 | 0 |
Related includes definitely, probably, or possibly related.
Percentages for the number of subjects with AEs/SAEs are based on the number of subjects in the safety population (N).
Subjects who experienced one or more adverse events in a category are counted only once.
Infections were more frequent with MMF maintenance compared to withdrawal (32 [64%] maintenance; 24 [46%] withdrawal; Table 2). Although most infections were mild, more severe infections were reported in the maintenance group (six in four maintenance participants; one in one withdrawal participant). The prevalence of infection was also greater with MMF maintenance (appendix p 20).
Discussion:
MMF is commonly used to treat SLE, particularly in those with severe kidney disease. However, MMF is teratogenic, highly immunosuppressive, and prolonged use is associated with substantial AEs, including serious infections.19,20 Therefore, discontinuing MMF is attractive in SLE patients whose disease activity has been suppressed and stabilized. However, the risks of increased disease activity with MMF withdrawal are poorly understood. In this trial, we found that MMF withdrawal is not significantly inferior to MMF maintenance in quiescent SLE patients on long-term MMF. We provide estimates for the rates of disease reactivation and flares and increases in risk with withdrawal to assist clinicians in making informed decisions on withdrawing MMF in stable patients. Balancing the risk of toxicities against increased disease activity is critical for informed therapeutic decisions.
Although MMF withdrawal was not significantly inferior to MMF maintenance, we report increases in risk with confidence limits to assist clinicians in making informed decisions on withdrawing MMF in stable patients. Specifically, we found a 6-8% increase with upper 85% CLs ranging from 11-19% in CSDR and flare risk with MMF withdrawal compared to maintenance. The 85% upper CL defines a range of values where the actual increase in risk for the population is likely to reside. Specifically, if the study were replicated repeatedly, the actual population value would be included in the 85% confidence range 85% of the time. Clinicians can be reasonably confident (85%) that the actual increases in risks for the population range from ≤11% (BILAG A) to ≤19% (any SELENA-SLEDAI flare). When considering MMF withdrawal, clinicians and patients must decide if these upper limits fall within personal levels of acceptable risk. Clinicians may weigh the risks of serious events more heavily and should balance the risks of reactivation and flares with those of long-term MMF use. Likewise, patients must also understand CLs, which requires thoughtfully prepared patient education materials and non-directive support by a trained coach.21,22
We also found that the increase in risk is higher and the CLs wider in individuals with a history of renal involvement, although not statistically different. A recent study found that lupus nephritis patients with low-grade histological activity are more likely than patients with normal biopsies to flare following the withdrawal of maintenance immunosuppression;23 therefore, it is possible that the increased CSDR risk in individuals with a history of renal disease in our study could be due to a subset of patients with residual low-grade histologic activity. Future studies are needed to test this hypothesis. However, our findings are consistent with the non-inferiority WIN-Lupus trial in proliferative lupus nephritis patients receiving maintenance azathioprine or MMF and HCQ.24 The WIN-Lupus trial demonstrated that a significantly higher proportion of patients in the discontinuation group experienced severe SLE flares and the time to severe SLE flare was shorter in the discontinuation group. In the WIN-Lupus trial, patients were required to be taking maintenance MMF or azathioprine for at least 2 years and no more than 3 years, which is less than the average MMF duration in our study (6.6 years) and could potentially account for the difference in significance between our and the WIN-Lupus study. Although the WIN-Lupus results are inconclusive, our two studies corroborate the finding of a modest increased risk with therapy withdrawal. Our study and the WIN-Lupus trial illustrate the challenges in evaluating the safety of withdrawing therapy in quiescent SLE populations. First is quantifying the meaning of “safe withdrawal,” which, for these studies, was the acceptable level of increased risk with therapy withdrawal. Because our team felt an acceptable increase in risk would have to be quite small to encompass widely-varying clinical circumstances, we opted for an estimation study without pre-specifying an acceptable risk level. Secondly, recruitment for our study slowed dramatically after the initial bolus of eligible patients, possibly because of our requirement for two years of quiescence for those with a history of renal involvement, ending with 102 of the target 120 participants. WIN-Lupus also recruited only 96 of its target 200 due to their strict inclusion criteria and patients’ pregnancy desires. Finally, both studies were open-label due to cost considerations. While unmasking allows for informed decisions on modifications to lupus medications, it could also lead to overestimating the risks for endpoints dependent on treatment changes, particularly in the withdrawal group. In our study, this was a concern for CSDR and SELENA-SLEDAI flares. However, we observed only three cases (two maintenance, one withdrawal) where the SELENA-SLEDAI flares associated with CSDR were based solely on meeting criteria for increases in immunosuppression medications. Several limitations must be considered when interpreting our results. In an effort to approximate a “real-world” study, our study was open-label and participants and staff were aware of treatment randomization, which could affect the results. Participants were required to maintain stable doses of HCQ for ≥12 weeks before and throughout the study. Thus, risks may differ for those not taking HCQ. Furthermore, some bias may have been introduced based on investigator selection of subjects who may have had differing exposure to MMF and/or more or less severe disease history. In addition, the mean duration of MMF use was ≥6 years. We have not investigated the relationship between risks and duration of disease quiescence on MMF therapy. In addition, most participants had lupus nephritis; therefore, a larger study including more patients without lupus nephritis is needed to provide more accurate risk profiling for both groups. Determining changes in risk of renal flare by kidney biopsy would also be valuable in lupus nephritis patients. Similarly, a longer duration, such as 3 years as observed in the ALMS study25 would be useful to better determine if MMF withdrawal reduces treatment toxicity without leading to an unacceptable risk of flare. Furthermore, MMF dose was relatively low in both study arms, suggesting that participants had already tolerated MMF reduction before enrollment; therefore, these findings may not translate to patients on high doses of MMF. We also did not measure HCQ and MMF levels, so adherence to treatment could not be directly assessed in this study, although we did perform pill counts for the MMF maintenance arm for the duration of the study. Finally, estimated risks for a target population do not inform us about risks to individuals. Studies to find risk factors and biomarkers associated with disease reactivation and flares would help us better understand personal risks associated with MMF withdrawal. Furthermore, these results are from patients with quiescent disease and an average of 6 years of MMF therapy, who investigators felt were appropriate for enrollment to this randomized trial. These results should not be considered appropriate to extrapolate to all SLE patients taking MMF.
Our findings suggest that MMF may be safely withdrawn in patients with stable SLE; however, larger studies with a longer follow-up are still needed. Our study is the first clinical trial of therapy withdrawal focusing on quiescent SLE patients, with or without lupus nephritis, on long-term MMF. Estimates for the rates of disease reactivation and flares and increases in risk with MMF withdrawal will assist clinicians and patients in making informed decisions on MMF discontinuation.
Supplementary Material
Research in context.
Evidence before this study
Mycophenolate mofetil (MMF) is commonly used to treat moderate/severe systemic lupus erythematosus (SLE) patients, particularly in those with lupus nephritis. However, long-term MMF treatment is associated with significant adverse events, including increased risks of infection, malignancies, impaired vaccine responses, severe birth defects, and miscarriage, demonstrating a need for MMF withdrawal following disease quiescence. However, the risks of MMF withdrawal remain unknown.
We searched PubMed for English publications using the search terms “therapy withdrawal” and “lupus” or “therapy weaning” and “lupus” and filtered the results by Randomized Controlled Trials, which yielded 22 results. Only one randomized controlled trial (WIN-Lupus) investigated MMF withdrawal in SLE patients. The study did not demonstrate non-inferiority of MMF or azathioprine withdrawal for time to relapse in lupus nephritis patients. Maintenance therapy withdrawal was associated with a higher risk of severe SLE flare. The WIN-Lupus trial was limited to lupus nephritis patients, a fifth of whom were receiving azathioprine and not MMF as maintenance immunosuppressive therapy.
Added value of this study
This study was a multi-center, open-label, randomized trial as part of the Autoimmunity Centers of Excellence Network analyzing the risks associated with MMF withdrawal compared to maintenance in patients with quiescent SLE on long-term MMF. While 17·6% (n=9) of patients who withdrew MMF had clinically significant disease reactivation (CSDR) compared to 10·2% (n=5) of those who maintained MMF by week 60, the risk of CSDR did not significantly differ between groups. Although the differences were not significant, this study used an estimation-based design to determine estimated increases in CSDR risk with 75%, 85%, and 95% confidence limits to assist clinicians and patients in making informed treatment decisions. Specifically, we found a 6-8% increase with upper 85% confidence limits of 11-19% in CSDR and flare risk following MMF withdrawal. While the proportion of patients who experienced an adverse event were similar between groups, infections were more frequent with MMF maintenance (32 [64%]) compared to withdrawal (24 [46%]), with a 6-fold increase in severe infections in the maintenance group (6 in 4 maintenance participants versus 1 in 1 withdrawal participant). To our knowledge, this is the first randomized trial that analyzed only MMF withdrawal in SLE patients with or without lupus nephritis.
Implications of all the available evidence
In SLE patients with quiescent disease on long-term MMF (an average of 6 years), MMF withdrawal is not significantly inferior to MMF maintenance. However, patients and clinicians should be informed of the slight increase in flare risk to aid treatment decisions.
Acknowledgments:
The authors would like to thank all study participants and study staff at all sites. We thank Joan Merrill, MD, and Catriona Wagner, PhD, for critical review of the manuscript.
Funding source:
The study was supported by awards from the Autoimmunity Centers of Excellence, a research network supported by the National Institute of Allergy and Infectious Disease (NIAID/NIH). Grant support from the National Institutes of Health includes (U19AI082714, UM1AI144292, P30AR073750 [JAJ, EFC]), and to Rho, the statistical and clinical coordinating center (HHSN272200900057C and 5UM2AI117870 [LKE, BB])
Conflicts of interest:
AS received consulting fees from AstraZeneca, AbbVie, GSK, Kezar Life Sciences, Eli Lilly, Bristol Myers Squibb and lecture fees from Astra Zeneca, Rheumatologic Dermatology Society, and Lupus Therapeutics. GC was a member of independent data monitoring boards of clinical trials by Roche and VERA therapeutics. BHR received consulting payments from Roche/Genentech, Aurinia, Novartis, Alexion, GSK, Kyverna, Kezar, HI-Bio, and Tome and served on the Board of Directors for NephroNet and the Scientific/Medical Advisory Board for the Lupus Foundation of America and Lupus ABC/LRA. RJL received support for the present manuscript as part of the Autoimmunity Centers of Excellence grant. DE received grants or contracts from ACR/EULAR, NIH, GSK, and Exagen, royalties or licenses from UptoDate, consulting fees from Chugai, AbbVie, and Argenx, and lecture fees from GSK and Aurinia. JM is an employee of DAIT/NIAID, the sponsor and funding agency of this trial. CA received a grant or contract from GSK, consulting fees from GSK, AstraZeneca, Bristol Myers Squibb, and AbbVie, and served on a data safety monitoring or advisory board for Alumis. LKE serves as a consultant for Rho Federal Division. All other authors declare no competing interests.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Data Sharing
Data are accessible from ImmPort (SDY2195; https://www.immport.org/shared/study/SDY2195).
References:
- 1.Yen EY, Singh RR. Lupus-An unrecognized leading cause of death in young females: A population based study using Nationwide Death Certificates, 2000-2015. Arthritis Rheumatol 2018;70:1251–1255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Davis LS, Reimold AM. Research and therapeutics-traditional and emerging therapies in systemic lupus erythematosus. Rheumatology (Oxford) 2017;56:i100–i113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wieske L, van Dam KPJ, Steenhuis M, et al. Humoral responses after second and third SARS-CoV-2 vaccination in patients with immune-mediated inflammatory disorders on immunosuppressants: a cohort study. Lancet Rheumatol 2022;4:e338–e350. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Food and Drug Administration. CellCept [package insert], Genentech USA, Inc; 2020: Southern San Francisco, CA. (https://www.accessdata.fda.gov/drugsatfda_docs/label/2018/050722s035.050723s035,050758s033,050759s041lbl.pdf) Accessed 10/22/2020. [Google Scholar]
- 5.Mycophenolate REMS (Risk Evaluation and Mitigation Strategy. (https://www.mycophenolaterems.com/.) Accessed 10/22/2020.
- 6.Hahn BH, McMahon MA, Wilkinson A, et al. American College of Rheumatology guidelines for screening, treatment, and management of lupus nephritis. Arthritis Care Res (Hoboken). 2012;64(6):797–808. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Fanouriakis A, Kostopoulou M, Cheema K, et al. 2019 Update of the Joint European League Against Rheumatism and European Renal Association-European Dialysis and Transplant Association (EULAR/ERA-EDTA) recommendations for the management of lupus nephritis. Ann Rheum Dis. 2020;79(6):713–23. [DOI] [PubMed] [Google Scholar]
- 8.Hochberg MC. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1997;40(9):1725. [DOI] [PubMed] [Google Scholar]
- 9.Buyon JP, Petri MA, Kim MY, et al. The effect of combined estrogen and progesterone hormone replacement therapy on disease activity in systemic lupus erythematosus: a randomized trial. Ann Intern Med 2005;142:953–62. [DOI] [PubMed] [Google Scholar]
- 10.Frane J. A method of biased coin randomization, its implementation, and its validation. Drug Information Journal 1998;32:423–432. [Google Scholar]
- 11.Gordon C, Sutcliffe N, Skan J, Stoll T, Isenberg DA. Definition and treatment of lupus flares measured by the BILAG index. Rheumatology (Oxford) 2003;42:1372–9. [DOI] [PubMed] [Google Scholar]
- 12.Petri M, Kim MY, Kalunian KC, et al. Combined oral contraceptives in women with systemic lupus erythematosus. N Engl J Med 2005;353:2550–8. [DOI] [PubMed] [Google Scholar]
- 13.National Cancer Institute, National Institutes of Health, U.S. Department of Health and Human Services. Common Terminology Criteria for Adverse Events (CTCAE) Version 4·0; NIH publication # 09-7473. Published May 29, 2009; Revised Version 4·03 June 14, 2010. [Google Scholar]
- 14.Castrejon I, Tani C, Jolly M, Huang A, Mosca M. Indices to assess patients with systemic lupus erythematosus in clinical trials, long-term observational studies, and clinical care. Clin Exp Rheumatol 2014;32:S85–95. [PubMed] [Google Scholar]
- 15.Montan I, Lowe B, Cella D, Mehnert A, Hinz A. General Population Norms for the Functional Assessment of Chronic Illness Therapy (FACIT)-Fatigue Scale. Value Health 2018;21:1313–1321. [DOI] [PubMed] [Google Scholar]
- 16.Fendl DM, Ware JE. Patient-reported functional health and well-being outcomes with drug therapy: a systematic review of randomized trials using the SF-36 health survey. Med Care 2014;52(5):439–45. [DOI] [PubMed] [Google Scholar]
- 17.Jolly M, Pickard AS, Wilke C, et al. Lupus-specific health outcome measures for US patients: the LupusQoL-US version. Ann Rheum Dis 2010:69(1): 29–33. [DOI] [PubMed] [Google Scholar]
- 18.Ibanez D, Urowitz MB, Gladman DD. Summarizing disease features over time: I. Adjusted mean SLEDAI derivation and application to an index of disease activity in lupus. J Rheumatol 2003;30(9):1977–82. [PubMed] [Google Scholar]
- 19.Singh JA, Hossain A, Kotb A, Wells G. Risk of serious infections with immunosuppressive drugs and glucocorticoids for lupus nephritis: a systematic review and network meta-analysis. BMC Med 2016;14(1): 137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Perez-Aytes A, Marin-Reina P, Boso V, Ledo A, Carey JC, Vento M. Mycophenolate mofetil embryopathy: A newly recognized teratogenic syndrome. Eur J Med Genet 2017;60:16–21. [DOI] [PubMed] [Google Scholar]
- 21.Rahn AC, Backus I, Fuest F, et al. Comprehension of confidence intervals -development and piloting of patient information materials for people with multiple sclerosis: qualitative study and pilot randomised controlled trial. BMC Med Inform Decis Mak 2016;16:122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Jull J, Köpke S, Smith M, et al. Decision coaching for people making healthcare decisions. Cochrane Database Syst Rev I2021;11:CD013385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Malvar A, Alberton V, Lococo B, et al. Kidney biopsy-based management of maintenance immunosuppression is safe and may ameliorate flare rate in lupus nephritis. Kidney Int 2020:97:156–162. [DOI] [PubMed] [Google Scholar]
- 24.Jourde-Chiche N, Costedoat-Chalumeau N, Baumstark K, et al. Weaning of maintenance immunosuppressive therapy in lupus nephritis (WIN-Lupus): results of a multicentre randomised controlled trial. Ann Rheum Dis 2022;81:1420–1427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Dooley MA, Jayne D, Ginzler EM, et al. Mycophenolate versus azathioprine as maintenance therapy for lupus nephritis. N Engl J Med. 2011;365(20):1886–95. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are accessible from ImmPort (SDY2195; https://www.immport.org/shared/study/SDY2195).